Kevin Staley - US grants

The training and research program described in this proposal supports the application for a Clinical Investigator Development Award for Dr. Kevin Staley. The proposed program will enable Dr. Staley to develop, under the supervision of Dr. David Prince, the skills necessary to pursue a research career in the area of pediatric epilepsy. The proposed research is motivated by the clinical difficulties involved in the treatment of neonatal seizures. The goal of this project is to evaluate two of the many possible hypotheses regarding the poor response of neonatal seizures to current anticonvulsant therapy. These hypotheses are based on the well-documented immaturity of the hippocampal GABAergic inhibitory system in the neonatal rat hippocampus, and the conditions under which GABAA receptor-mediated functions have been shown to be decreased. 1) Is modulation of GABAA inhibition by barbiturates and benzodiazepines different in the neonatal vs adult hippocampus? 2) Are the effects of the barbiturates and benzodiazepines on the GABAA system minimized under conditions which are likely to occur during neonatal seizures: depletion of presynaptic GABA, alteration in transmembrane ionic gradients, accumulation of intracellular free calcium, and the depletion of intracellular high-energy phosphates? The research will focus on the modulation of GABAA receptor-mediated inhibition in areas CA1 and CA3 of the in vitro hippocampal slice preparation, and will utilize the whole-cell patch clamp recording technique. Experiments include 1) Measurement of the effects of barbiturates and benzodiazepines on GABAA receptor-mediated evoked and spontaneous synaptic events; 2) Determination of the effects of barbiturates and benzodiazepines on GABAA synaptic events under conditions which are likely to occur during neonatal seizures 3) Assessment of alternative modulators of GABAergic inhibition such as GABA- aminotransferase inhibitors and steroids.

DESCRIPTION: This proposal is based on the hypothesis that the neuronal membrane depolarization mediated by intense activation of inhibitory GABAA receptors plays an important role in epileptogenesis, so that selectively modulating the ionic basis of this GABAergic depolarization represents a unique anticonvulsant strategy. The rationale for this hypothesis is based on the recent demonstration of a mechanism of activity-dependent GABAergic membrane depolarization: intense dendritic GABAA receptor activation results in differential collapse of the opposing transmembrane concentration gradients of HCO3 and Cl-, the anions that permeate the GABAA ionophore. The resulting shift of the GABAA reversal potential depolarizes the dendrites and enables excitatory NMDA receptor activation by diminishing the voltage- dependent Mg2+ block of the NMDA ionophore. Thus epileptogenic alterations in the balance between excitation and inhibition may occur due to activity-dependent shifts in anionic concentration gradients that change the function of GABAA receptors from inhibition to excitation. To test the hypothesis, the applicant will use whole-cell and extracellular recordings in the hippocampal slice preparation to 1) measure the transport rates of the anions that participate in the GABAergic depolarization 2) test how the modulation of those transport systems affects the activity-dependent GABAergic depolarization 3) test how the modulation of the ion transport systems affects epileptogenesis in acute in vitro seizure models. These experiments may provide alternative therapeutic strategies for seizures that are resistant to treatment with GABAergic anticonvulsants, including the barbiturates and benzodiazepines. Such strategies should prove particularly useful for neonatal seizures, since activity- dependent GABAergic neuronal excitation is especially prominent in the immature CNS, and the clinically available anticonvulsant acetazolamide has been shown in the applicant's preliminary studies to selectively block the GABAergic depolarization.

These experiments will test the hypothesis that changing the driving force for the synaptic currents gated by activation of the inhibitory GABAA receptor is an effective way to increase the efficacy of these currents and thereby increase neuronal inhibition. The rationale for this hypothesis is based on Ohm's law: the amount of current that flows through inhibitory GABAA channels can be altered either by changing the GABAA receptor-mediated conductance, or by changing the driving force for the current. Although many available drugs increase the GABAA receptor-mediated conductance, the driving force has not been manipulated. This is because the processes that establish the anionic transmembrane equilibrium, and thus the driving force for GABAA currents, are just now being understood at a molecular level. One determinant of the neuronal anionic transmembrane equilibrium is CIC-2, a cloned inward rectifier that stabilizes the Cl- equilibrium potential near the resting membrane potential. To test our hypothesis, adenovirus vectors will be used to express CIC-2 in cultured sensory neurons and in granule cell neurons of the hippocampal dentate gyms. These neurons maintain high intracellular Cl- levels, are depolarized rather than hyperpolarized by GABAA receptor activation, and do not naturally express ClC-2. The impact of CIC-2 on Cl- homeostasis, the efficacy of GABAergic inhibition, and potentiation of inhibition by anticonvulsants will then be tested at the cellular level using whole-cell recordings, and at the network level by measuring input-output responses of the dentate gyrus. These experiments will determine whether the effects of anticonvulsant drugs that increase the conductance through which inhibitory synaptic currents flow can be complimented by increasing the driving force for those currents. This new approach to the modulation of neuronal inhibition may provide a foundation for the development of alternate strategies for the treatment of intractable epilepsy.

DESCRIPTION (provided by applicant): The goal of this proposal is to understand the interactions between the activity of a neural network and the properties of the synapses that link the individual neurons. The rationale for the proposed studies is based on several discoveries made during the last grant cycle. In area CA3 of the disinhibited hippocampal slice preparation, bursts of network activity are possible when there is sufficient strength of the recurrent collateral synapses that link the pyramidal cell in this area. Thus short-term, activity-dependent depression of the strength of these synapases terminates bursts, and recovery from this short-term depression determines when the next burst is possible. Further, long-term determinants of the strength of these synapses also affect burst probability, so that both long-term potentiation (LTP) and long-term depression (LTD) produce corresponding long-term changes in burst timing. Under the appropriate conditions, burst activity can induce either LTP or LTD. In the case of LTD, the changes in strength are stabilized by an associated change in the threshold for further LTP or LTD induction. In the next cycle we propose to test two hypotheses related to these findings. First, one component of short-term synaptic depression in CA3 appears to be mediated by dendritic calcium influx. We will test the hypothesis that adenosine acting at presynaptic A1 receptors mediates this unusual mechanism of short-term depression. Second, we will investigate the mechanism by which the threshold for long-term synaptic plasticity is altered in CA3. We will test the hypothesis that a reduction in the maximum NMDA receptor-mediated postsynaptic calcium influx is associated with LTD. Finally, we will test how these relationships between synaptic properties and network behavior may affect epileptogenesis and the treatment of seizures. These investigations will help us understand what determines the timing of interictal epileptic activity, which is needed in order to use the temporal pattern of interictal activity to predict the probability of future seizures. In addition, these investigations will help us understand how to produce stable, long-term decreases in the probability of seizures by selective induction of LTD at the most active synapses in epileptic foci.

DESCRIPTION (provided by applicant): Excitation spreads through a neural network via positive feedback connections between the neurons. The amount of positive feedback in the network is determined by the number and strength of these excitatory synaptic connections, as well as the degree to which these connections are masked by pre and postsynaptic inhibition. The proposed research will test the hypothesis that the amount of positive feedback in a neural network is correlated with the probability that the network will initiate a seizure. To test this hypothesis, we have developed two noninvasive methods. The first method quantifies the amount of positive feedback based on the temporal pattern of interictal spikes on the electroencephalogram (EEG). The second method modifies the amount of positive feedback by selective long-term depression (LTD) of the strength of recurrent excitatory synapses. Using a well-characterized rat kainate model of chronic epilepsy, the amount of positive feedback measured from the EEG will be correlated with seizure probability during epileptogenesis. As an additional test of the hypothesis, the amount of positive feedback in the epileptic networks will be decreased by LTD of the recurrent synapses, and the seizure probability will be compared to EEG measures of positive feedback before and after LTD. These experiments may provide two important tools for treating epilepsy. The first is the ability to estimate seizure probability from the pattern of interictal spike activity on the EEG, which would make possible the prospective evaluation of the risk of seizures and the efficacy of anticonvulsant therapy. The second is the induction of long-term decreases in seizure probability by synapse-specific LTD of recurrent excitatory synapses in the epileptic network, which may prove to be a very useful anticonvulsant strategy.

The GABAA receptor is unique in that its activation can either excite or inhibit neurons. The ability of the GABAA receptor to switch between excitation and inhibition is a key feature of a surprising number of neural processes, including neocortical development, thalamocortical oscillations, circadian rhythm generation, the response of nerve cells to both trauma and prolonged seizures, and the perception of pain. Variations in GABAA function within a given population may occur both acutely and over more protracted periods of time. The dual effect of GABA in these systems is made possible by changes in the reversal potential for Cl-, the ion that carries the majority of the GABAA receptor-mediated postsynaptic current. Neurons alter the Cl- reversal potential (E(Cl)) by changing the intracellular Cl- concentration [Cl-]in. Neuronal Cl- homeostasis is mediated by a variety of passive and active mechanisms such as anion exchange, voltage-gated anion channels, and cation-Cl- cotransporters. In particular, the family of cation-Cl- cotransporters has the capability to efficiently move Cl- into or out of cells depending on which specific cotransporter predominates. In the peripheral and central nervous systems specific cation-Cl cotransport proteins have been identified: KCC1 and KCC2 move Cl- out of the cytoplasm while NKCC-2 moves Cl- into the cytoplasm. This proposal is designed to investigate the hypothesis that changes in intracellular Cl initiated by a variety of processes are primarily effected by two molecular mechanisms: expression and post-translational regulation of neuronal Cl transport gene products. We propose to first test the hypothesis that the regulation of [Cl-]in is due to the balanced expression of inwardly and outwardly directed Cl- transport proteins. We will then examine how specific second-messenger signal transduction pathways regulate the direction and rate of Cl- transport. Discovering how the family of cation-Cl- cotransporter genes maintains [Cl-]in and thereby determines whether GABAA activation is excitatory or inhibitory can lead to the identification of novel and highly specific therapeutic strategies, which will have special application to a broad spectrum of pathologies, including the treatment of seizures, modulation of sleep, amelioration of pain and spasticity, and may give new insight into neuronal responses to neurotrauma.

DESCRIPTION (provided by applicant): Electroencephalographically detected interictal spikes have played a central role in the diagnosis of epilepsy for over 60 years. Despite the importance of interictal spikes in the diagnosis of epilepsy, we don't know why interictal spikes are present in epileptic patients, nor do we understand the mechanisms that determine spike timing. Three findings from our experiments in the hippocampal CAS in vitro model of interictal spikes form the rationale for the proposed research: 1) interictal spikes induce long-term potentiation of the strength of excitatory synapses in the epileptic network that generates the spikes. 2) the temporal pattern of interictal spikes is in turn dependent on the strength of the excitatory synapses in the epileptic network. 3) partly blocking NMDA receptor-mediated calcium influx during interictal activity results in long-term depression rather than potentiation of the strength of the excitatory synapses in epileptic networks. We hypothesize that in vivo, 1) the temporal pattern of interictal activity reflects the strength of the excitatory synapses between neurons in the epileptic network. 2) interictal spikes are associated with epilepsy because they drive the strengthening of excitatory synapses in the epileptic network 3) long-term decreases in seizure risk can be produced by transient, partial NMDA receptor blockade during interictal spike activity to induce long-term depression of the strength of the excitatory synapses in the epileptic network. To test these hypotheses in vivo, we will correlate interictal spike patterns with seizure risk using newly developed techniques for continuous EEG radiotelemetric monitoring and computer analysis of the EEG data in a rat model of chronic epilepsy. We will use acute in vivo and chronic organotypic slice preparations to determine the maximum amount of LTD that can be induced using the NMDA antagonist technique, and then use this technique to induce long-term reductions in seizure risk in vivo. These experiments will 1) provide the basis for using EEG data to quantitatively determine long-term seizure risk, and 2) develop a powerful and noninvasive technique for inducing long-term reductions in seizure risk without the need for daily medication.

DESCRIPTION (provided by applicant): This research is directed at two conditions in which seizures respond poorly to anticonvulsant therapy: neonatal seizures and intractable temporal lobe epilepsy (TLE). These two conditions share an unusual feature: in both conditions, neurons are excited rather than inhibited by GABA, the principal inhibitory neurotransmitter. GABA is excitatory because neurons in these conditions accumulate intracellular chloride, which reverses GABAA receptor-mediated current flow. NKCC1 is a chloride transporter that imports chloride into neurons, and KCC2 is a chloride transporter that exports chloride. In the proposed research we will test whether NKCC1 activity exceeds KCC2 activity in the neonate and in the kainate model of TLE. Transporter function and expression will be evaluated in the neonatal brain, in the kainate model of TLE, and in adult control animals. We will use whole-cell and gramicidin perforated patch recordings in hippocampal slices to measure the kinetics of these two transporters. We will use western blots and immunocytochemistry to determine the level of expression of the two transporters. The NKCC1 chloride transporter is exquisitely sensitive to the diuretic bumetanide. Using a kainate model of acute neonatal seizures in the rat pup, and the kainate model of chronic TLE in the adult rat, we will test whether blocking NKCC1-mediated chloride accumulation with bumetanide will restore GABAA receptor- mediated inhibition and thereby ameliorate these two types of seizures. To quantify the effects on seizures we will use acute and chronic, radiotelemetric digital EEG recordings and computerized seizure analysis. Bumetanide has already been tested as a diuretic in human neonates at doses that inhibit NKCC1, so bumetanide treatment of neonatal seizures is a new and feasible treatment of a disorder for which there is no effective therapy and a very high risk of lifelong morbidity. Similarly, bumetanide-induced restoration of GABAA receptor-mediated inhibition in intractable TLE could provide a non-surgical therapeutic alternative to patients who are not candidates for epilepsy surgery due to the location of their ictal onset zones.

[unreadable] DESCRIPTION (provided by applicant): We are requesting NINDS support for a first-time Gordon Research Conference on Mechanisms of epilepsy and neuronal synchronization to be held Aug 6-11, 2006 at Colby College in Maine. The main goal of the study of seizures is to identify the mechanisms underlying synchronous electrical discharges in neuronal networks in order to develop more effective and less toxic treatments and cures for epilepsy. A unique, intellectually challenging aspect of epilepsy research arises from the fact that it encompasses virtually all major levels of biological organization, from genes and ion channels to circuits and behavior. The major purpose of this Gordon conference is to bring together geneticists, molecular biologists, developmental neuroscientists, neuroanatomists, electrophysiologists and computational neuroscientists working on basic mechanisms related either directly or indirectly to seizure generation to synthesize current advances and to set the stage for future discoveries. Topics to be covered include Gene expression profiling in epilepsy, Epileptogenic ion channels, Epileptogenic dysgenesis, Homeostatic plasticity, Structural reorganization in epilepsy, Hypersynchrony and non-linear dynamics, Gap junctions, Key events underlying epileptogenesis, and Entorhinal-hippocampal interplay in epilepsy. Our goals are to disseminate the latest scientific advances, foster productive new insights and collaborations, and set the stage for new translational studies that will bring the newest discoveries to the bedside in the shortest possible time. [unreadable] [unreadable] [unreadable]

DESCRIPTION (provided by applicant): Interictal spikes on the EEG are strongly correlated with a propensity for seizures, but we don't understand why. In this application, we will test the hypothesis that interictal spikes occur as a consequence of monophasic spread of excitation through the epileptic network, while seizures are initiated when excitation spreads via unique, circular paths of excitation. These circular paths engender repeated, reentrant activation of the neural network. To use a cardiac analogy, reentrant activation of the epileptic network arises instead of a spike in the same manner that reentrant cardiac tachyarrhythmias arise instead of a QRS complex. This hypothesis has three testable components 1) there are multiple paths by which excitation may spread from a stochastic site of spike onset to the rest of the epileptic network. 2) Which path is followed depends on interactions between the onset site and changes in network circuitry induced by dysgenesis, injury, and transient local refractoriness 3) Some of these excitation pathways are closed loops, engendering re-entrant waves of excitation that underlie the local rhythmic activity recorded at the start of focal-onset seizures. We will test these hypotheses using in vivo and in vitro electrical and optical recordings of the spread of excitation in neural networks, as well as computer modeling. This project will provide insights into several pressing problems. First, we will better understand how focal seizures start, which may improve our ability to detect and abort them. Second, we may solve the puzzle of the relationship between spikes and seizures. Third, we will be able to study the effect of anticonvulsants on reentrant activity, which may represent the earliest phase of a seizure;this would comprise a new and potentially more informative screen for drug efficacy. Fourth, in analogy to management of cardiac arrhythmias, this research provides the foundation for abortive stimulation or very focal ablation of the specific neural pathways that initiate seizures, making possible less invasive treatment of drug-resistant epilepsy. PUBLIC HEALTH RELEVANCE: One of the most disabling aspects of epilepsy is the unpredictable nature of seizures. This research will help us understand how epileptic seizures start, so that we will be able to predict and abort them. We will also understand how seizures are related to the activity observed on EEGs obtained between seizures. We predict that seizures start by activation of very specific neuronal pathways that can be ablated by surgery;such surgery should be as minimally invasive as the treatment of cardiac arrhythmias by cardiac catheterization. This would enable treatment of seizures arising in areas of the brain subserving language, movement and vision.

DESCRIPTION (provided by applicant): During the last cycle of this grant, we established that block of inwardly-directed cation-chloride co-transport alters the neuronal steady-state chloride concentration (Cl-i) in developing neurons, and consequently improves the efficacy of GABA-mediated synaptic inhibition and the control of seizures in the newborn. This finding forms the basis of new clinical trials in the US and Europe. We also found that cation-Cl transporters are at equilibrium at resting Cl-i. However, we don't know how to reconcile this finding with data from experiments utilizing the Cl-sensitive fluorescent dual wavelength fluorophore, Clomeleon, which demonstrate that each neuron has a unique Cl-i that is quite different from its neighbors. The question we address here is: how can cation-Cl transporters be at equilibrium at so many different Cl-i? The classic view that equilibrium depends only on Cl and cation concentrations does not explain the variance in Cl-i. Neuronal cation-chloride transporters obligately move water with cations and Cl, so that these transporters move isotonic (135 mM) cation-Cl solution into and out of neurons. Thus neuronal cation-Cl transporters also transport cytoplasmic volume, which alters the cytoplasmic hydrostatic pressure. This predicts that transmembrane hydrostatic and osmotic pressure gradients contribute to the free energy of cation-Cl transport and thus the equilibrium Cl-i. For example, a neuron with lots of osmotically active protein should have a lower equilibrium Cl-i than a neuron with less protein. Our primary hypothesis is that the pressure gradient across the neuronal membrane contributes to the free energy of transport and thus the Cl-i at which transport is at equilibrium. This hypothesis has important implications for prolonged seizures, which induce changes in the neuronal cytoskeleton that increase the volume of neurons, thereby lowering the osmotic pressure. Thus a linked secondary hypothesis is that seizure-induced changes in osmotic pressure favor movement of cations, Cl-i and water into neurons via cotransporters so that Cl-i increases and GABA signaling becomes excitatory. We will test these hypotheses by measuring neuronal volume and Cl-i at steady state and in response to ionic and osmotic challenges, seizures, and specific transport inhibitors. We will use mice that genetically express Clomeleon, acute &organotypic slice preparations, in vitro and in vivo multiphoton microscopy, pH-sensitive dyes, electrophysiological recordings, and transporter phosphorylation studies. Sensitivity of cation-Cl transport to local pressure gradients would allow neurons to maintain the incredibly precise geometries needed for stable cable properties and connectivity despite dynamic subcellular and intercellular fluctuations in osmotically active protein content. The hypotheses also predict that clinically available diuretics and inhibitors of cytoskeletal changes might also be useful in the treatment of status epilepticus in both developing and mature nervous systems. PUBLIC HEALTH RELEVANCE: Prolonged epileptic seizures damage the cytoskeleton of cortical neurons, allowing the neurons to expand and accumulate the salt and water needed to fill the new intracellular space. One component of salt is chloride, and its accumulation causes the neurotransmitter GABA to promote rather than inhibit seizures. We will investigate the mechanisms of chloride accumulation and strategies to prevent it in order to improve the treatment of prolonged seizures.

DESCRIPTION (provided by applicant): Feedback inhibition is the process by which activity in principal neurons stimulates interneurons to release the inhibitory neurotransmitter GABA onto the active principal neurons to prevent runaway excitation. Failure of feedback inhibition is thus a critical element in most theories of the pathogenesis of seizures. However, the functional anatomy of feedback inhibition in the normal brain and epileptic focus is unknown. Recent developments in optogenetics and multiphoton microscopy have made it possible to address this question directly. Here we propose to combine channelrhodopsin-mediated photoactivation of targeted pyramidal cells with high-speed multiphoton imaging of the responses in interneurons expressing the calcium fluorophore yellow chameleon 3.6. These techniques will allow us to define the location of the interneurons that are activated by the target pyramidal cell. After brain injury, the loss of principal neurons and interneurons is compensated by sprouting of new synaptic connections. Our computer modeling suggests that the circuit complexity engendered by this sprouting leaves the interneurons vulnerable to activity-induced synaptic depression that permits runaway excitation and seizures. We hypothesize therefore that epileptic circuits will be defined by characteristic changes in the anatomy of feedback inhibition: more principal neurons will share the same interneuron feedback networks, and individual interneurons will be activated by a wider anatomical range of pyramidal cells, so that the anatomical complexity of local feedback circuits will be increased in epileptic foci. We will test this hypothesis directly with the new optogenetic and microscopy tools in vitro using chronically epileptic organotypic slice cultures, and in vivo using chronically epileptic animals. This data will provide a critical new insight into the pathophysiology of epilepsy that we have not been able to acquire despite wonderfully detailed electrophysiological and classical anatomical studies. Testing for characteristic circuit alterations in epilepsy will make possible new classes of therapeutic interventions including pharmacological manipulation of activity-dependent depression, as well as preemptive activation of critical circuit elements.

DESCRIPTION (provided by applicant): Rapid advances in developmental neuroscience, imaging, genetics, and molecular therapeutics have thrust Child Neurology into the biomedical limelight. As a child neurology residency program director I am amazed at the increasingly spectacular qualifications of each new wave of resident applicants who are drawn to the fundamental problems facing our specialty. At Massachusetts General Hospital we are poised with a talented pool of residents interested in research, an abundance of basic science advances waiting to be translated to medical realities, an exceptionally rich research environment, and a PI, steering committee, and group of carefully selected mentors who together will guide the development of these trainees into successful physician scientists. We have assembled a group of accomplished NIH-funded mentors whose expertise is focused on brain development and neurological diseases of childhood. These are experienced investigators with a strong track record of training physician scientists. The K12 participants will have full access not only to their mentor's lab but also the many resources available through collaborative programs designed to accelerate neuroscience research at Massachusetts General Hospital, Harvard Medical School, Harvard School of Public Health, and the Massachusetts Institute of Technology. A carefully structured curriculum of coursework, written evaluations, and meetings with the PI and child neurology co-mentors has been carefully designed to ascertain that all scholar participants advance smoothly to independent research funding.

DESCRIPTION (provided by applicant): It has been extremely difficult to screen for drugs that prevent epilepsy after brain injury, because spontaneous recurrent seizures begin gradually and the interval between seizures varies widely, so that long-term, intensive seizure monitoring is required to determine whether a drug prevents epilepsy. A rigorous yet rapid screen would enable us to test many promising compounds for anti-epileptogenic properties, and would provide epilepsy researchers with a much-needed assay to test new compounds developed within their laboratories. We have developed a new set of technologies to enable large-scale screening for antiepileptic drugs. In collaboration with the Harvard Center for Engineering in Medicine, we have developed an in vitro model of epileptogenesis comprised of glass chips that can be used to culture, record from, and administer drugs to arrays of 32 organotypic brain slices per chip. We have recently shown that these brain slice cultures undergo a rapid, predictable process of epileptogenesis, and that the organotypic brain slices respond to anticonvulsant drugs just as human patients do. We have recently published computer algorithms for continuously recording and quantifying electrographic spikes and seizures, and in collaboration with Ed Dudek, an investigator in the National Institute of Neurological Diseases and Stroke (NINDS) Anticonvulsant Screening Program, we have validated these algorithms in two in vivo models of epileptogenesis as well as in the in vitro model. We will use our newly-created technologies to execute large-scale, parallel screening of drug libraries for agents that prevent, reduce, or reverse epileptogenesis. One such library is the NINDS Custom Compounds library; we will focus on the 561 compounds that have already been approved by the FDA. We propose to subject these compounds to a rapid, rigorous two stage screen for anti-epileptogenic properties. In the first stage, we will screen compounds using parallel cultured brain slice assays. Compounds that prevent, reduce, or reverse epileptogenesis will then progress to the second stage, in which the most promising compounds will be subjected to a more rigorous albeit much slower second assay, the kainate model of epileptogenesis. In both stages, electrographic seizure activity will be assayed quantitatively using continuously recorded and analyzed EEG data. This research will lead to a U01. The results of the R21-funded screening project may produce compounds that are sufficiently active to begin clinical testing. Alternatively, the R21 will provide lead compounds that will enable us to apply for the NINDS Blueprint Grand Challenge for New Drugs for Diseases and Disorders of the Nervous System, so that we can screen larger compound libraries and optimize compounds in collaboration with the medicinal chemists at The Laboratory for Drug Discovery in Neurodegeneration (LDDN) in the Harvard NeuroDiscovery Center. PUBLIC HEALTH RELEVANCE: It has not been possible to screen for drugs that prevent epilepsy after brain injury, because seizures begin to occur very gradually, and the interval between seizures varies widely, so that long-term, intensive seizure monitoring is required to determine whether a drug prevents epilepsy. We have developed a set of technologies that makes possible the application of large-scale screening strategies to this problem. Cultured brain slices that become spontaneously epileptic will be deployed in a highly parallel screening program as a first stage in screening. In vivo testing will be used to confirm the effects of the most promising agents found in the first stage of screening.

DESCRIPTION (provided by applicant): The NINDS has issued U- and R-series RFAs this year to discover drugs to prevent the development of epilepsy after brain injury. The rate-limiting factors in the response to the RFAs are the lack of highly scalable models of epilepsy after brain injury, and the means to quantify epileptogenesis. The hippocampal slice is a compelling model of traumatic brain injury, and organotypic cultures of these slices provide a unique window into recovery after severe shear injury. We re-discovered the utility of the organotypic slice culture as an accelerated model of post traumatic epileptogenesis 20 years after epileptic activity in this preparation was first described. The epilepsy research community lost 20 years of use of this exceptionally powerful model largely because inadequate standardization created a perception of inadequate reproducibility. We propose to optimize the organotypic slice preparation for the study of epileptogenesis, focusing on 3 key questions to maximize its utility: First, what are the best means to quantify epileptogenesis in this preparation? We have developed methods for continuous electrographic recordings and automated seizure detection algorithms that make feasible the parallel processing of arrays of epileptogenic slice cultures. We will determine whether the methods that we have devised for quantification of in vivo epileptogenesis can be applied to these culture arrays. Second, what are the optimum culture conditions for epileptogenesis in slice cultures? This preparation was never optimized for the study of epilepsy, so we need to ascertain that metabolic substrates and products do not limit the quantity of epileptic activity. Third, what surrogate markers can be used to accelerate the use of this system for screening? Although quantification of epileptogenesis from electrographic recordings is the gold standard, our preliminary data suggest that there are potentially promising markers such as extracellular lactate levels and intracellular chloride and calcium levels that can be used as the basis of fluorescence assays that will substantially accelerate throughput with this model. Together, these studies will enable a new, rapid, highly scalable, readily transferrable, and quantifiable assay of post traumatic epileptogenesis that will accelerate both fundamental research into pathogenesis and the search for therapeutic strategies. PUBLIC HEALTH RELEVANCE: Organotypic brain slice cultures are a potentially powerful model of post traumatic epilepsy, but this preparation was never optimized for the study of epilepsy. We have the experience with both the culture and seizure quantification necessary to develop this preparation to the point that it can become a rapid, highly scalable, reproducible and transferrable, and quantifiable model of post traumatic epilepsy. Such a model will profoundly increase our capacity to seek new cures for epilepsy after brain injury.

Seizures are named for their unpredictability. This project seeks to understand the conditions that lead to seizure initiation so that we can develop anticonvulsant strategies to ameliorate those conditions. Organotypic hippocampal slice cultures first develop interictal spikes and then spontaneous seizures over 3 - 7 days in vitro (DIV). The question addressed here is: why does a seizure occur instead of another interictal spike? Answering this requires a careful dissection of the conditions just prior to each spike, and each seizure. That is not feasible to do even in vitro, whereas in computer models, pre-ictal conditions can be stored, analyzed, manipulated, and re-run. In the last round, we developed the first neural network model that generates both interictal spikes and seizures. The seizures are self-sustaining reentrant waves of activity that have recently been recorded with high-density electrodes in human epilepsy. The conditions for sustained re-entry are difficult to achieve, which may account for the relative rarity of seizures vs. interictal spikes. In this proposal, we will test three pre-ictal conditions that we have identified in silico. The first condition is the wiring of the network. Neuronal synaptic wiring strategies are largely unknown, but we have developed technologies to determine synaptic wiring in the organotypic slice, including a microcscope within an incubator (?IncuScope?) and deployment of digital micromirror chips to stimulate single neurons using optogenetics (?optical synapses?). We will read out the synaptic targets of the stimulated neurons using sensitive new transgenic calcium fluorophores. We will then test the necessity of that ictal wiring pattern by altering it using excitatory and inhibitory optical synapses. The second pre-ictal condition is the pattern of refractoriness in the network. These patterns shape the wave of activation, and most waves die out without forming stable re- entrant cycles. The third preictal condition is the location of the onset of the wave of activation; reentrant waves of activity can only be initiated from particular regions within a pre-ictal network pattern of refractoriness. We will test the second and third conditions by converting spiking networks to seizing networks, and seizing networks to spiking or quiescent networks, with appropriate stimulation using excitatory optical synapses. These experiments will provide unique new insights into brain connectomics; ictogenesis; and the mechanisms of efficacy and failure of the therapies for medically intractable epilepsy: seizure surgery (changing network wiring; Aim 1) and brain stimulation (changing the refractory patterns in epileptic networks; Aim 2).

Summary The latent period between brain injury and subsequent epilepsy is months to years in duration, providing a unique window for antiepileptogenic therapy. Recent reports of promising experimental disease-modifying therapies can't advance to clinical trials until three hurdles are overcome: First, quantification of epilepsy is necessary to assess efficacy of interventions, but epilepsy after brain injury is very difficult to quantify: the latency to first seizure is long and variable, and early seizures are often subtle, infrequent, and clustered between long inter-cluster intervals. Second, because of the long latency between injury and epilepsy, clinical trials need to be quite prolonged and therefore prohibitively expensive. Third, because ~ 20% of moderately brain-injured patients develop epilepsy, most of the treated patients could not benefit from long-term antiepileptogenic therapy, despite exposure to the risks and side effects. These three hurdles could be overcome with sufficiently accurate biomarkers. We recently demonstrated that early electrographic epileptiform activity is a promising predictor of epilepsy after brain injury induced by kainic acid. Here we propose to address critical knowledge gaps regarding electrographic biomarkers of epileptogenesis. The predictive power of electrographic biomarkers has not been assessed after more clinically relevant injuries such as trauma and hypoxic-ischemic injury. The predictive power of electrographic biomarkers has not been systematically compared to the predictive power of traditional physical descriptors of injury, such as lesion size and location. Furthermore, it has not been determined whether combining electrographic and physical-injury parameters would improve their predictive power. We will employ well-established models of clinical injuries, the lateral fluid percussion (LFP) trauma model and the Rice- Vannucci model of focal hypoxia-ischemia in P30 rats. The incidence of epilepsy in these models is close to the human experience, and thus provides a more rigorous test of the predictive power of these biomarkers than the kainate model. Further, the latency to seizures is sufficiently long in these models to enable testing as to whether the appearance of early electrographic biomarkers is more closely related to the time elapsed since the injury, or to the time remaining prior to the first seizure; the nature of these relationships will significantly impact the design of clinical studies of these biomarkers. In Aim 1, we will use a novel miniature telemetry device for continuous recording of video-EEG together with validated, unbiased computer detection algorithms to quantify early epileptiform activity and seizures in these brain injury models. We will optimize EEG sampling and develop the best predictive model based on epileptiform electrographic activity and injury descriptors, and then prospectively test this model in a second group of animals. In Aim 2, we will use the same approach to test whether early electrographic epileptiform activity and injury descriptors predict the severity of epilepsy, including latency to first seizure and seizure frequency, once epilepsy is fully developed.

? DESCRIPTION (provided by applicant): Dramatic new insights into the functioning of neural networks have been made possible by our ability to visualize neural function with calcium-sensitive fluorophores, and biology has been revolutionized by the ability to sequence and manipulate DNA, RNA, and proteins. Both of these tremendous advances have unexplored flip sides. Our understanding of neural network function remains limited by our inability see GABA-mediated synaptic activity: we can't measure the output of the remarkable diversity of interneuron structure and function. Similarly, the methods for studying templated biopolymers such as DNA, RNA, and protein are not applicable to untemplated biopolymers such as polyglutamylated intracellular tubulin, and the variably sulfated glycosaminoglycans (GAGs) that comprise the extracellular matrix. The glutamate and sulfate moieties displace chloride, thereby defining chloride microdomains. These microdomains thus provide a photographic negative of the information that may be stored in local untemplated biopolymers. Several compelling lines of evidence suggest that chloride microdomains alter the direction and magnitude of currents flowing through open GABAA receptors, providing a mechanism to read out information stored in the anion distribution of the biopolymers. This implies a unique direction and size of chloride current at each GABAA receptor, so that these critical determinants must be measured to understand the contribution of interneurons to network operation. To visualize chloride microdomains, we have assembled a uniquely qualified team to create and fully characterize a series of reporters of chloride concentration placed at the most critical location: the intra and extracellular faces of GABAA receptor subunits. Super Clomeleon, a new ratiometric chloride- sensitive fluorophore, will be fused to g2 and d subunits of the GABAA receptor. g2-linked Super Clomeleon will report synaptic chloride microdomains that drive synaptic GABA signaling, while d-linked Super Clomeleon will report chloride domains subserving extrasynaptic GABA signaling. Both intra and extracellular chloride microdomains will be analyzed by the Super Clomeleon - GABAA subunit fusion constructs, so that the direction and amplitude of chloride currents can be precisely mapped. The GABAA subunit-fluorophore fusion products will be parallel-processed in four labs so that in the three- year time window of this RFA we can assess receptor assembly, membrane trafficking, receptor conductance, chloride sensitivities, gains, pH dependence, and stability in vitro and in vivo. We can then make timely go / no go decisions regarding the utility of going forward to create transgenic mice.

? DESCRIPTION (provided by applicant): In the last grant cycle, we characterized, validated and published a unique in vitro tool for the study of epileptogenesis - hippocampal organotypic slice cultures that develop spontaneous seizures after 1 week in vitro. We used this preparation to blindly screen over 500 drug-concentration combinations for activity in chronic, post-traumatic epilepsy at speeds that are orders of magnitude faster than any other therapeutic testing strategy for chronic epilepsy. We then used double-blind in vivo EEG testing in the kainate model of chronic epilepsy to confirm the anticonvulsant effect of a lead compound, celecoxib. A key innovation that made these speeds possible was the use of lactate and lactate dehydrogenase (LDH) levels in spent media as assays for seizure burden and cell death, respectively. In searching for the cause of the increase in lactate, we found a persistent neuronal membrane leak that increases cytoplasmic sodium and calcium (? Nai & Cai) days before histochemical evidence of cell death. New data indicate that COX2 induction leads to translocation of Bax, a canonical mitochondrial permeabilizing protein, to the cytoplasmic membrane, where it forms pores that admit Na+ and Ca2+. We propose to test the following pathophysiology: traumatic or ictal injury induces Ca2+-dependent Bax translocation to the cytosolic membrane, where it creates a progressive Ca2+ and Na+ leak that should kill the neuron. However, neurons survive for some time due to their uniquely high ion transport capacity. The ion transport consumes a lot of ATP, and lactate production is a consequence of ATP generation. Progression of the leak eventually leads to membrane depolarization, which may contribute to ictogenesis, and cell death. We will test these ideas by correlating seizures, lactate, and ATP production in Aim 1. In Aim 2 we will establish the nature of the membrane leak. In Aim 3 we will evaluate the consequences of the membrane leak on ATP production, membrane potential, ictogenesis, and cell death. We propose to use cell-type specific expression of ratiometric, fluorescent reporters of Na+, Ca2+, ATP, NADH, lactate, caspase and membrane potential in the organotypic slice model, together with multiphoton and custom- built low-light, wide-field microscopes to address these questions at temporal and spatial resolutions that have not previously been feasible.

? DESCRIPTION (provided by applicant): The two major inhibitory synaptic transmitters, GABA and glycine, gate anion currents whose direction and magnitude are determined primarily by the distribution of chloride ions (Cl-) on either side of the neuronal membrane. This Cl- distribution s surprisingly difficult to determine. Each neuron seems to have a unique reversal potential for currents gated by GABAA receptors (EGABA), and recently EGABA has been shown to vary at different subcellular regions of the same neuron. These findings are hard to reconcile with the idea that one of two equilibrative cation-Cl cotransporters determine the Cl- distribution, because such a scheme links the Cl- distribution to the K and Na distributions, and should lead to an extremely uniform and predictable EGABA. ++ In the last grant cycle, we found a candidate explanation for the variability of EGABA: Cl is displaced by anionic - macromolecules in the cytoplasm and extracellular space. This is congruent with the concentration of impermeant anions in the cytoplasm, and the density of variably-sulfated glycosaminoglycans that largely comprise the matrix filling the extracellular space. The distribution of Cl- by charge displacement is analogous to the distribution of Styrofoam packing peanuts in a shipping box, where the shipped contents are the anionic macromolecules, the box is the neuronal membrane, and the peanuts are Cl-. Cl- distribution by displacement has interesting and testable predictions that we will begin to explore here. First, Cl- microdomains would be created if the distribution of anionic macromolecules is not uniform. Intracellular Cl microdomains would alter EGABA locally, while extracellular Cl domains would primarily affect -- the local GABAA conductance. Conceivably then, every GABAA synapse could have a unique reversal potential and conductance, which would permit the read-out of the enormous amount of information that could be stored in the distribution of Cl-displacing anionic macromolecules. The second prediction is that disruption of the extracellular matrix after brain injury, for example by the activation of matrix metalloproteases, would increase extracellular Cl-. Equilibrative co-transport of Cl-, cations, and water would then increase intracellular volume - a new mechanism for cytotoxic edema. We will test these predictions using perforated patch recordings and novel high-resolution Cl- reporting tools. These tools include transgenic mice with inducible expression of more sensitive ratiometric fluorescent Cl- fluorophores. We are also developing, as part of a collaborative BRAIN U01, fusions of these new Cl- fluorophores to the intracellular and extracellular faces of GABAA receptors. These fluorophores provide the sensitivity, stability and spatial resolution to rigorously test the Cl- microdomain hypothesis and the pathogenesis of cytotoxic edema using multi-photon and newly-developed very-long-term imaging technologies.

Abstract The NINDS Anticonvulsant Screening Program (ASP) has identified most of the anticonvulsants in clinical use today. However, one third of epileptic patients do not respond to these drugs. The ASP protocols are based on seizures induced by subjecting normal animals to acute convulsant conditions. We have developed a complimentary system of novel in vitro and in vivo assays of spontaneous seizures in chronically epileptic preparations. This two-stage screening system provides a unique focus on recurrent spontaneous seizures in chronic epilepsy models. The first stage is an in vitro assay comprised of the organotypic hippocampal slice culture, which develops electrographic seizure activity and corresponding biochemical biomarkers over the first week in vitro. The second stage is an in vivo assay comprised of the kainate model of epilepsy in which spontaneous seizures are monitored using continuous telemetry and supervised, blinded, computerized seizure detection. We used the rapid in vitro assay to screen over 400 compound-concentration combinations from the NINDS Custom Compound Collection. We found a lead compound, celecoxib, and then verified this lead by the second-stage testing in a randomized double blind in vivo crossover trial. Celecoxib had no effect on seizures induced by acute application of convulsants to normal brain tissue, suggesting that its anticonvulsant properties are unique to chronic epilepsy, and raising the possibility that its spectrum of action will be distinct from anticonvulsants discovered by the ASP protocols. The next step in development is medicinal chemistry to optimize celecoxib?s anticonvulsant efficacy. This is most feasibly accomplished through the UH2 / UH3 Blueprint Neurotherapeutics Network. As our discussions with BPN program officers clarified, to efficiently utilize the BPN medicinal chemistry program we must further develop the in vitro and in vivo assays and acquire additional data on our lead compound. The UH2/3 mechanism was considered the most appropriate funding mechanism by the NINDS program officer. In the R21 phase of this proposal, we will extend the in vitro assay?s concentration-response for celecoxib and 2,5 dimethyl celecoxib, a derivative that does not inhibit COX2 but has equal anticonvulsant efficacy in vitro. We will then characterize the assay?s reproducibility and Z factor. We will also establish the dose-response of the in vivo assays for celecoxib, and increase the number of in vitro and in vivo sites to two each in order to improve robustness and throughput, as well as engage outstanding younger investigators in this effort. In the R33 phase of the proposal, we will further characterize the lead compound by determining whether COX2 inhibition is necessary for anticonvulsant activity in the in vitro and in vivo assays.